Leveraging screenomics to identify mental illness: Detecting bipolar disorder through computational analysis of smartphone screen data

Mental illnesses like bipolar disorder affect millions of people around the world, but early symptoms are often difficult to detect. Working across the disciplines of clinical psychology, communication, and computer science, my research will develop a novel computational tool to identify signals of mania and depression in real-time. I will apply cutting-edge computational analytics developed in the Human Screenome Project to unobtrusively examine moment-by-moment changes in people’s smartphone behavior to flag warning signs of mental illness with (1) a massive existing dataset of 700 individuals digital behavior and mental health, and (2) a clinical cohort of 50 patients with bipolar.